Abstract Human leukocyte antigen (HLA) class I molecules have profound influences on infectious disease and cancer outcomes, via effects on immunity mediated by CD8+ T cells and natural killer (NK) cells. Three sets of genes, HLA-A, HLA-B and HLA-C encode HLA class I proteins. These genes are among the most polymorphic of human genes, with thousands of alleles found in humans. Each allotype presents a unique set of peptides at the cell surface, and individual peptides in complex with HLA class I molecules confer exquisite specificity for recognition by T cell receptors (TCR) of CD8+ T cells. In the canonical textbook-defined HLA class I assembly pathway, peptides that bind to HLA class I molecules are typically derived from the cytoplasm of cells, and transported into the endoplasmic reticulum (ER) by the transporter associated with antigen processing (TAP). In the ER, peptides assemble with HLA class I molecules in a process that is guided by specific assembly factors such as tapasin and generic ER chaperones. Peptide-loaded versions of HLA class I molecules exit the ER, whereas ER quality control is thought to retrieve peptide-deficient HLA class I for degradation, due to their relative instability. A number of our recent findings indicate that the canonical assembly pathway does not fully account for cell surface expression patterns of HLA-B allotypes in all cells, and additionally that peptide- deficient (empty) HLA-B do exist at the cell surface under some conditions, and are functional in the immune response. Based on these findings, it is our central hypothesis that HLA-B allotypes are functionally separable based on the stabilities of their empty forms and their peptide-binding preferences, and that these characteristics determine their competence for non-canonical assembly pathways and related functions. To test this hypothesis, we examine the model that HLA-B allotypes vary both in their constitutive cell surface expression levels and in cross-presentation efficiencies in antigen presenting cells (APC), based on competence for non-canonical assembly. We also examine the model that HLA-B allotypes vary in their abilities to induce effective CD8+ T cell immunity against Epstein Barr Virus (EBV) infections due to varying competencies for assembly in a TAP-deficient environment. Finally, the models of differential induction and novel functions for empty HLA-B in CD8+ T cell and NK cell functions are examined. Together, these studies address the central idea that the extreme polymorphisms of the HLA-B locus, which are evolutionarily-selected mutations within a confined region of the HLA class I structure, generate a hierarchy of protein folding and assembly phenotypes that are exploited by distinct arms of the immune response to maintain multi- compartmental and multimodal surveillance. These studies are expected to guide our progress in precision medicine by identifying best candidate HLA-B for specific vaccine targets, and provide new targets for combatting HLA-B-driven autoimmunity and drug hypersensitivity.